Forwarding this because it apparently didn't make it to the list. Begin forwarded message:
From: David Carlisle <davidc@nag.co.uk> Date: September 17, 2010 8:44:51 AM PDT To: webkit-dev@lists.webkit.org Cc: mjs@apple.com Subject: Re: [webkit-dev] HTML5 & MathML3 entities
Not sure I can post to this list (hence cc to Maciej) but replying to the thread from July
http://www.mail-archive.com/webkit-dev@lists.webkit.org/msg11986.html
there were a couple of points raised.
* that rang changed definition
and
* entities expanding to more than once character not being supported.
On the first:
rang did not used to be (directly) a CJK character it was in the math symbol block at 2329 but erroneously given a canonical decomposition to the CJK block at 3008. Unicode recognised the error but never change canonical decompositions so they deprecated 2320 and introduced a new character 27E8 that is the same apart from this decomposition.
http://www.unicode.org/charts/PDF/U2300.pdf
explicitly says these characters are deprecated:
Also, even if Unicode had not deprecated 2329, the W3C "Unicode Normal Form" bans entities using characters that have canonical decompositions as it makes entity expansion and NFC canonicalisation interact badly. Changing the definition is unfortunate but sort of inevitable given the history.
On the second: It's probably worth noting that currently the HTML5 spec only lists entities that expand to a single character, but there is an open bug entry about that:
http://www.w3.org/Bugs/Public/show_bug.cgi?id=10067
David
________________________________________________________________________ The Numerical Algorithms Group Ltd is a company registered in England and Wales with company number 1249803. The registered office is: Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.
This e-mail has been scanned for all viruses by Star. The service is powered by MessageLabs. ________________________________________________________________________